Sie sind auf Seite 1von 17

Revitalizing Production Logging

Thousands of high-angle and


horizontal wells have been
drilled in the last ten years.
Steve Bamforth
BP Exploration Operating Co. Ltd.
Poole, England
Christian Besson
Ken Stephenson
Colin Whittaker
Cambridge, England
George Brown
BP Exploration Operating Co. Ltd.
Sunbury on Thames, England
Grard Catala
Gilles Rouault
Bernard Thron
Clamart, France

44

Gilbert Conort
Montrouge, France
Chris Lenn
Dubai, United Arab Emirates

As a result, there are many


mature fields with complex
well production problems.

Brad Roscoe
Ridgefield, Connecticut, USA

Today, new technology and

For help in preparation of this article, thanks to Ashok


Belani, Schlumberger Wireline & Testing, Montrouge,
France; John Ferguson, Schlumberger Cambridge
Research, Cambridge, England; Yves Manin,
Schlumberger Riboud Product Center, Clamart, France;
Jean-Rmy Olesen, Beijing, China; DeWayne Schnorr,
Anchorage, Alaska, USA; Antonio Jorge Torre, Technical
Editing Services, Houston, Texas, USA; and Amal
Vittachi, GeoQuest, Dallas, Texas.
BorFlo, CPLT (Combinable Production Logging Tool),
FloView, FloView Plus, PLT (Production Logging Tool),
PL Flagship, PVL (Phase Velocity Log), RST (Reservoir
Saturation Tool), TDT (Thermal Decay Time) and
WFL (Water Flow Log) are marks of Schlumberger.

flow in wellbores have

better understanding of fluid

revived production logging


methods for all types of wells.

Oilfield Review




,,,,,




,,,,,





,




,,,,,


,,,



,,

,,,






Fault

Stagnant gas

Gas

Failed external
casing packer

Formation
instability

Oil layer

Cuttings

Fractures

Water

For decades, production logs have been


used in new wells to optimize ultimate
recovery and to help avoid potential production problems. In older wells, these logs
aid in diagnosing declining production and
planning remedial work.1
From the outset, production logging (PL)
has been used to determine the dynamic patterns of flow rates of water, oil and gas under
stable producing or injecting conditions by
answering the following questions: How
much of the well is flowing? Which zones are
producing oil, water and gas? How much of
each type of fluid is flowing from each zone?
Ideally, PL techniques should identify each
fluid, measure the volume fraction of each
fluid in the pipecalled the holdupand
its velocity, and from these compute flow
rates.2 Traditional PL measurements use turbine flowmeters called spinners for velocity,
gradiomanometers for density, capacitance
for holdup, manometers for pressure and

Winter 1996

thermometers for temperature. Of these five


measurements, only velocity and density
tend to be used in traditional quantitative
PL analysis.
The reliability of the data generated by
traditional PL logging depends almost
exclusively on the type of well being
logged. In vertical wells with high flow
ratesusually from 200 to 5000 B/D [30 to
800 m3/d], depending on the tool used and
the pipe diameterthese PL measurements
and their analysis usually produce reliable
results. However, in some wells, phenomena such as flow behind casing or interzone flow make traditional PL difficult.
The upsurge in deviated and horizontal
wells creates boreholes with very different
fluid flow characteristics from vertical wells,
adding further complexity to multiphase
flow and radically changing the physics and
technology of fluid-flow measurement
(above). In gas-and-liquid or oil-and-water
flow, the lighter phase moves rapidly along
the high side of the borehole, establishing a

The challenges facing production logging in horizontal wells. Trapped fluids


can directly affect production and influence the data from a production log, especially sensors such as spinners and capacitance tools. Because horizontal wells
inevitably have doglegs and undulations,
stagnant water may lie either inside or outside the casing in low areas at the bottom
of the well; stagnant gas may accumulate
on the high side of drainhole undulations.
These nonflowing fluids distort measurements. Changes in the flowing cross-sectional area have a direct impact on spinner response (inset, left). Horizontal wells
are frequently completed uncemented,
using prepacked screens or slotted liners
with external casing packers (ECPs). An
ECP that fails to set properly or formation
collapse create volume changes that
affect flow velocities. Faults, fractures and
formation instabilities may cause fluid
crossflow. Cuttings on the low side of the
borehole may alter fluid velocities and
result in erroneous readings.

circulating current that often causes a backflow along the lower side (see Fluid Flow
Fundamentals, page 61).
Depending on the borehole deviation, the
velocity and holdup of the different phases
can change dramatically for any given flow
rate. In these circumstances, traditional PL
measurements may become unreliable. 3
This article looks at how new techniques are
helping to shed light on flow in complex
vertical wells, and to deliver PL measurements in deviated and horizontal wells.
1. Wade RT, Cantrell RC, Poupon A and Moulin J: Production Logging (The Key to Optimum Well Performance, Journal of Petroleum Technology 17 (February 1965): 137-144.
2. For an authoritative treatment of multiphase flow: Hill
AD: Production Logging-Theoretical and Interpretative Elements, SPE Monograph 14, 1990.
3. Brown G: Using Production-Log Data From Horizontal Wells, Transactions of the SPWLA 36th Annual
Logging Symposium, Paris, June 26-29, 1995, paper SS.

45

When to Run Production Logs

Generally, PL has two important applications: measuring well performance with


respect to reservoir dynamics and analyzing
mechanical problems in the borehole.
Although decisions to run production logs
usually depend on specific reservoir economics, there are general guidelines.
First, PL may be used in new wells to evaluate initial production and verify the
integrity of the completionfor example,
indicating where there is flow behind casing. When initial performance does not
meet expectations, information from PL may
often point to remedial work to optimize
production and suggest different completion
techniques for future wells.
A special use of PL in horizontal, high-rate
wells is to verify friction-induced production
loss in long drainholes. This friction loss
sometimes negates any extra productivity
expected from the long drainhole, and a
better choice would be to drill multiple,
shorter lateral sections in a stacked or fanshaped pattern.4
Second, PL should be considered for any
well that shows sudden decreases in production or increases in gas/oil ratio (GOR)
or water cut.
Third, just as a yearly checkup by a physician is prudent, PL may be used periodically
to detect problems such as water or gas coning, or fingering before extensive production
loss occurs. This is particularly important for
dump-flood wells, where PL is the only
monitoring method.5
Fourth, injection wells may be initially
analyzed and then monitored with PL.
Knowledge of where injected fluids are
going is critical for avoiding undesired
flooding that leads to serious problems such
as casing-annulus crossflow, the creation of
unswept and trapped hydrocarbons, and
water-wet damaged formations.

X25

Zone

X50

1:200 m

Radius of Bit
0

10

GR (C.H.)-GR (O.H.)
10
0

Openhole Sw
1993
100
p.u.

Casing Wall
Assumed Cement
Sheath
Formation
Perforated Zone
Scales

Openhole Sw
1996

Borehole Water
Borehole Oil

Openhole Porosity
50

100

p.u.

Openhole Porosity

Openhole Fluid Volume


1993
50

p.u.

p.u.

RST Fluid Volume


1996
50

p.u.

50

p.u.

Shale 1
Shale 2
Feldspar
Quartz
Calcite

Downhole Flow Rate


0
B/D 10000

RST Oil 1996

Water

RST Oil
1996
Nonmovable Oil (O.H.)

Water
Oil
Gas

Openhole CPLT-RST evaluation from South China Sea. Track 1 (left) contains a well
sketch with casing (black) and a cemented casing-formation annulus (gray hatching).
Uranium scale was indicated by the difference in natural gamma ray activity between
the openhole and most recent cased-hole gamma ray survey. Track 2 contains the openhole log and the latest RST water saturation analysis. Track 3 shows the production logs
and static-fluid volume analysis in the formation. The top of Zone 3 at X41 and the top
half of Zone 2 at X47 still shows some unproduced oil. Zones 1 and 4 are completely
depleted. The production logs shows most of the water production coming from the top of
Zone 2 at X46 m.

46

Oilfield Review

The ability to carry out downhole PL measurements in a stabilized well under


dynamic conditions is the key to successful
production management. The resulting
downhole flow-rate determination may be
compared with stabilized surface flow rates.
This quantitative comparison between
downhole and surface flow rates allows
detection of any surface-to-downhole discrepancies caused by such factors as tubing
leaks, thief zones, unwanted fluid entries or
other hydraulic malfunctions.
Production Logging in Vertical Wells

Increasingly, operators incorporate PL into


their reservoir monitoring programs. Today,
this often includes cased-hole saturation
logging techniquessuch as thermal-neutron decay time or carbon-oxygen measurementsrun in combination with traditional
PL tools to provide an enhanced understanding of reservoir dynamics.6
The RST Reservoir Saturation Tool can be
used to make a snapshot of reservoir saturation. Repeating these measurements over
time helps monitor changes in reservoir saturation. But the dynamic description of
flow conditions obtained from production
log profiles is absolutely necessary to
unravel complex commingled production
in a many wells.
For example, to gain a clear picture of production dynamics in a declining reservoir, the
CPLT Combinable Production Logging Tool
log and the RST technique were used in combination in a reservoir located in the Pearl
River Mouth basin in the South China Sea.
The reservoir, a sand-shale sequence, was
producing from four commingled sandstone formations, and the operator needed
to understand current reservoir production
on a layer-by-layer basis. The CPLT-RST
reservoir monitoring suite was deployed in

Winter 1996

a well located at the top of the reservoir


(previous page). Openhole well evaluations, with the latest hydrocarbon volume
from RST C/O monitoring, showed the
changes in reservoir saturations.
The lowest zone had been completely
depleted, as had about half of the next
zone. A cased-hole versus openhole gamma
ray comparison revealed evidence of substantial scale buildup in the lowest perforated zones. This indicated that large volumes of water had been produced from the
lower zones, and scale could potentially
plug perforations.
The production logs provided the key to
understanding what was happening in the
well. The flowmeter and gradiomanometer
profiles showed that there was only a little
fluid production, mostly water, coming from
the lowest perforations. About 60% of the
total water production came from the second lowest set of perforations, and most of
that from just 2 m [6.5 ft] of the upper section of perforations.
Surprisingly, the RST monitor log indicated
that water production was coming from a
fully oil-bearing part of the formation. It was
suspected that the water was coning up
from the bottom part of the zone, now completely depleted of hydrocarbons. Logs from
other wells, downdip in the reservoir, confirmed this conclusion. Reducing the drawdown pressures may allow production of the
bypassed hydrocarbons, still contained in
this zone, to continue.
In the wells second highest perforated
zone, the RST monitor logs showed a significant oil-water contact (OWC). The lowest
half of the zone was fully depleted, whereas
the upper half was untouched by production. Unexpectedly, production log profiles
indicated greater hydrocarbon production
than water, perhaps because scale had
plugged the lower perforations in the
watered-out part of the zone. The upper perforations in this zone did not appear to be
plugged by scale, yet the production profiles

showed minimal contribution over the


entire interval. This result confirmed the
diagnosis from RST monitoring logs that the
upper formation layer had been swept of all
movable hydrocarbons.
Another example, this time in a vertical
well with a thief zone and borehole water
entry, occurred in Indias offshore Bombay
High field, operated by Indian Oil and Natural Gas Commission (ONGC). The reservoir was under waterflood, and the operator
needed to identify zones of water entry and
to determine whether flow was occurring
behind the casing. It was also suspected
that injection water had broken through
and was being produced from one of five
sets of perforations.
A WFL Water Flow Log tool was combined with the PLT Production Logging Tool
log to distinguish between flow inside and
outside the casing (see Fluid-Flow Logging
Using Time-of-Flight, page 50). The downhole flow rates were complex. The top of
the lowest set of perforations, Zone 5, produced only small quantities of water. There
was a large increase in water flow coming
from the second lowest set of perforations.
A modest amount of oil, 400 BOPD
[63 m 3/d], was also produced from this
zone. The middle set of perforations, Zone
3, also produced 1000 BWPD [160 m3/d]
with only a small amount of oil. The second
highest set of perforations showed no fluid
production (next page).
4. Hill D, Neme E, Ehlig-Economides C and Mollinedo
M: Reentry Drilling Gives New Life to Aging Fields,
Oilfield Review 8, no. 3 (Autumn 1996): 4-17.
5. In dump-flood wells, water is produced from an
aquifer and injected into a producing formation in the
same well.
6. Albertin, I, Darling, H, Mahdavi, M, Plasek R, Cedeo
I, Hemingway J, Richter P, Markley M, Olesen J-R,
Roscoe B and Zeng W: The Many Facets of Pulsed
Neutron Cased Hole Logging, Oilfield Review 8,
no. 2 (Summer 1996): 28-41.
An essential input for RST-A C/O monitoring logging is
the oil holdup in the borehole. The PL gradiomanometer provides this measurement.

47

200 GR

Gas
Oil

160

Water

200

60
API

Reconstructed Reconstructed
Fluid Velocity Fluid Density
ft/min
gm/cm3
0

125

0.0

1:200

125

6000.0

1500 Near

125

125

243

245

WFL
Water Flow Rate
BWPD
0.0

6000.0

80
40

80

0
0

900

24

36

48

60

48

60

0
0

300

12

40

600

Zone

0
12
Time, sec

120

120

1200

Measured
Measured Fluid Measured Fluid Temperature
3
F
Velocity ft/min Density gm/cm

Far

160

Count rate, cps

GR
0

Downhole
Flow Rate
B/D

12

24

24

36

36

48

48

60

60

X370

50

GR

40

30

X380

200 Far

20
160
10
120

0
X390

80
500

Count rate, cps

Thief zone in vertical well. The PLT-WFL interpretation


analysis indicated that Zone 1 is removing more than
120 BOPD and 2200 BWPD from the well. Crossflow had
been set up by the injection and production schemes.
At X354, the WFL decay-time distributions showed a
flow rate over 2000 BWPD inside the casing (inset,
above right). At X393 m, the WFL decay-time distributions showed that no flow was detected (inset, right).

400

Near

12

24

36

40

0
0

12

24

48

36

60

300
200
100
0
0
12
Time, sec

With the top set of perforationsZone 1


the picture changed dramatically. Here,
more than half of the production from the
four zones below disappeared into the formation. Zone 1 was acting as a major thief
zone, consuming 120 BOPD [19 m3/d] and
about 2200 BWPD [350 m 3/d] from the
well. This unusual crossflow, verified by
WFL results, indicates a pressure differential
between the two formation layers, which
was not present when the well was initially
put on production. The WFL survey also
indicated that there was no channeling
behind the casing.
Armed with this knowledge, the operator
had two choices for remediationsqueeze
the perforations in the lowest zones (3 to 5)
to prevent water production, or isolate
Zones 1 and 2 using a dual-completion
scheme, putting the long string on gas lift,
and allowing continued production of 400
BOPD [64 m3/d] from Zone 4.

48

24

Nonvertical Production Logging

Once a well substantially deviates from


vertical and multiphase flow becomes
complex, spinner tools often indicate only
reverse flowespecially when the spinner
is not centralized in the borehole, but lying
near the bottom where the reverse flow is
found (next page, right).7 Capacitance tools
may also measure the lower, denser phase
of the fluid giving misleading holdup data.
As the wells angle increases to horizontal,
flow becomes entirely stratified, and the
averaged mixture velocity from a flowmeter
spinner alone is meaningless.

36

48

60

Other phenomena affect PL measurements


in deviated and horizontal wells. For example stagnant fluids may confuse sensors;
fractures and faults may allow crossflow;
and failed external packers may introduce
variable flow regimes (see page 45).
Horizontal and many deviated wells are
often completed either open hole, with
uncemented slotted liners or with
prepacked screens. 8 Such completions
introduce other special fluid-flow and production problems that usually are not
encountered in vertical, cased wellssuch
as flow restrictions due to the logging tool in
the pipe forcing fluids to channel through
the liner-formation annulus. Furthermore, a

Oilfield Review

Flow Outside Pipe

Count rate, cps

400

240

Total count rate


Background

160
80

12

24

36

Time, sec

Flow Inside Pipe


2200

Count rate, cps

Water flow
Gas flow

Velocity = 8.5 ft/min


Rate = 439 BWPD

320


,,,,


,,,,



,,,,



,,,,

,,,,



48

60

Velocity = 8.8 ft/min


Rate = 850 BWPD

1760
1320

Total count rate


Background

880
440
0
12

24

36

Time, sec

48

60

Distinguishing between water flow inside


and outside casing. Time-of-flight gamma
ray time-decay distributions indicated
whether the flow is inside or outside the
casing. The lower graph shows the
response when water is flowing inside the
casing. The blue shaded area reflects the
final time-decay response to flowing water
after the background and standing water
signals have been removed. The blue area
had a sharply peaked response, which
indicated that the slug of activated water
flow occurred in a smooth cross-sectional
pipe area without dispersion. The top
graph indicates the magnitude and shape
of the time-decay response when flow is
outside casing. Here the time distribution
was much broader, reflecting slug dispersion as it flowed around the outside of casing. Lower total counting rates are due to
gamma ray attenuation in the casing.

special problem occurs near the uphole end


of a slotted liner. Here, annular fluids are
forced out of the annulus back into the liner
or casing, resulting in significant turbulence
that tends to mix the fluids. This turbulence
can encourage backflow to develop on the
low side of the hole, which can seriously
affect flowmeter readings.
In horizontal wells completed with conventional cemented liners, flowmeter spinner profiles look more like their vertical
counterparts, often showing smooth, distinct
evenly-separated profiles when recorded at
different speeds.9 However, cementing in
horizontal wells is usually not as successful

Winter 1996

as in vertical wells because the liner is


decentralized within the borehole, often
leading to cement voids and channels with
accompanying annular production.
Other problems in horizontal completions
include acceleration of fluids due to gravity
when undulations in the well profile are sufficiently large. If peaks of the flowmeter measurements are taken as representative of the
full mixture velocity, the trend is an increase
in velocity where the well turns downward
and a decrease as the flow reaches the
trough of the undulation. Backflow always
appears to occur in inverted, undulating
wells where the heavy phase falls down the
low side of the drainhole. In many cases, the
heavy phase (usually water) simply circulates
in the sump and is not produced.

Delivering Data from Deviated Wells

Success in isolating crossflow problems in


the offshore Bombay well convinced the
operator to try a combined WFL-PLT
approach in a cased-hole, deviated well that
was producing oil, water and gas. The operator was unsure of the exact location of the
water entry zones and whether these could
be sealed off using cement squeezes to
reduce water cut.

Backflow as drainhole moves towards


vertical. In highly
deviated or horizontal wells and at low
fluid velocities,
buoyancy forces
tend to segregate
fluids. The lighter
phase flows in the
upper part of the
pipe dragging
along with it some
of the heavier
phase. Sometimes
part of the heavier
phase moves downwards due to gravity, causing a circulation within the
pipe. Badly centralized flowmeters in
the lower portion of
the deviated pipe
will respond to this
downward flow.

Again, channeling behind casing was suspected. This time, the WFL measurements
showed this, and confirmed the PLT measurements in a difficult environment. The spinner
tool data below X050 indicated downflow,
the temperature gradient suggested possible
upward fluid movement and the gradiomanometer tool showed a single-phase
fluid below X050a very confusing picture.
The spinner measurement was presumed
unreliable in this zone, as it had insufficient
resolution to measure low apparent flow.
The thermometer was affected by fluid
movement inside and outside the casing,
but could not differentiate between the two
flow regions. The WFL data helped resolve
the dilemma, by distinguishing between
flows inside and outside the casing (above
left). In this case, water was flowing outside
(continued on page 52)
7. In this article, the range of deviated wells will include
moderate to the so-called high angle 30 to 85
from vertical; horizontal wells range from 85 to 95.
8. Brown G, reference 3.
9. Spinner turn rates are calibrated by logging at different
cable speeds.

49

Fluid-Flow Logging Using Time-of-Flight

Several years ago, the WFL Water Flow Log tech-

Near count
rate

nique was introduced using the TDT-P Thermal

Far count
rate

GR count
rate

Decay Time tool to provide water-velocity data,


first in vertical wells, then later in deviated and
horizontal wells.1 Today, the RST Reservoir Saturation Tool log provides water-velocity information
with more precision.2 A burst of fast neutrons from
Casing

the RST tool activates oxygen atoms in a small


Minitron

region surrounding the neutron source in the tool.

Oil

This includes any oxygen in the water flowing in


the pipe. Oil does not contain oxygen and there-

Water

fore is not affected. Activated oxygen atoms, in a


process like fluorescence, give off radiation, in the
form of gamma rays, radiating for a short time
after the neutron burst.
Moving water in the pipe will carry a cloud of
activated oxygen with it past the detectors in the

WFL Water Flow Log Measurements. A short burst of neutrons activates oxygen
in the surrounding water, and flowing water carries the activated cloud at the water
velocity. Source-detector distances and time-of-flight are used to determine the
water velocity.

tool (above right). The time between the neutron


burst and the detection of the activated water cloud
will be a time-of-flight for the water flow in the

Marker signal

Near detector borehole sigma indicator

pipe, and is used to compute water velocity. The


half-life of the oxygen activation is only seven seconds, so after a few minutes, the activation radiation has subsided to an undetectable level, making
the measurement environmentally safe.
There are two detectors in the RST tool.The tool
can use a variable neutron burst width from 0.1 to

10

20

30

3 sec with delays from 3.5 to 20 sec to measure


water-flow rates from as low as 6 ft/min [1.8 m/min]
to as high as 500 ft/min [152 m/min]. The RST tool

Oil-miscible marker

40
50
Time, sec

60

70

80

90

RST tool
Casing

may be inverted to measure downward water flow.


Oil

An additional gamma ray (GR) detector may be


incorporated in the logging tool string to measure
higher velocities.

Water

The RST-WFL technique may be used to measure other parameters. The total activation count

PVL Phase Velocity Log sonde

rate is proportional to the volume of water activated by the neutron burst, and therefore is a measure of the water holdup in the pipe. The time profile, or shape, of the activation count rate
distribution carries information about whether the

PVL Phase Velocity Log technique. A slug of oil-miscible marker fluid is injected
into the flowing oil phase, and is detected by the RST tool. The time-of-flight
between injection and slug detection along with the distance between the injector
tool and RST detector gives the oil velocity. The same process is used for water
phase-velocity measurements except a water-miscible marker compound is injected
into the heavier phase.

activated water is flowing near the tool in the borehole or behind the casing pipe in the annulus.

50

Oilfield Review

Borehole sigma indicator

144

300
490 ft/min
200
ft/min
ft/min

142
140

100
ft/min

Raw Data
Filtered Data

50
ft/min

12 ft/min

138
136
134
132
130

Oil
Water

200
0

10

15

20

25

60

80

100

120

750 BOPD

Time, sec
100

Typical marker slug time-of-flight distributions for a variety of fluid-flow velocities.


0

For horizontal wells, fluid flows are stratified,


500

with the light phase moving rapidly in the upflow

cause large changes in fluid holdup and the velocities of different phases, making it necessary to
know all fluid velocities. Spinners are usually not
applicable in stratified flow, and radioactive tracers are useful useful only for water-velocity measurements, because there are no oil-miscible
forms available. Radioactive tracers also have
strict procurement and safety issues.
The PVL Phase Velocity log also uses a time-offlight method to measure both oil and water velocities.3 This technique uses a chemical marker that
is injected into either the oil or water stream. The
time the marker takes to reach the detector is a
measure of fluid velocity (previous page, bottom).

1500 BOPD

400
100
300
0
200

100

0
0

100

200

300

400

500

Velocity, ft/min

borehole. Slight changes in borehole deviation

200

PVL measurements, ft/min

sections of the well along the high side of the

200

2300 BOPD

100

Actual water velocity, ft/min

PVL water velocity measurements in the flow-loop.


Water velocity measurements made using the PVL
technique for horizontal stratified two-phase flow (oil
and water), where the water holdup was kept at 50%,
show good agreement with actual controlled flow
rates. The error bars are dominated by the sampling
frequency of the borehole absorption measurement.

200

Flow-loop experiments at Schlumberger Cam-

200

3000 BOPD

100

The chemical marker contains a high concentration of the element gadolinium, which has a large
thermal neutron absorption cross section. The RST

bridge Research, Cambridge, England have vali-

tool senses the large increase in the borehole

dated the PVL measurements under a large variety

absorption cross section caused by the passage of

of flow conditions. Both single-phase oil and water

the gadolinium slug (above).

measurements show excellent agreement between

A high concentration of gadolinium chloride

PVL-measured and actual flow rates (above). Two-

[GdCl3] in water is used as a water-miscible

phase measurements, using oil and water or gas

marker. It has the high density and low viscosity

and water, demonstrate the ability to measure sep-

necessary for the water-phase measurements. For

arately each phase in a segregated flow (right).

the oil-phase measurements, a new, gadolinium-

1. Lenn C, Kimminau S and Young P: Logging of Water


Mass Entry in Deviated Well Oil/Water Flows, paper
SPE 26449, presented at the 68th SPE Annual Technical Conference and Exhibition, Houston, Texas, USA,
October 3-6, 1993.
2. Albertin et al, reference 6, main text.
3. Roscoe BA and Lenn C: Oil and Water-velocity Logging in Horizontal Wells Using Chemical Markers,
paper SPE 37153, presented the 1996 SPE International Conference on Horizontal Well Technology,
Calgary, Alberta, Canada, November 18-20, 1996.

rich compound, with low density and viscosity is


used. These markers are safe to handle, even in
concentrated form, and pose no environmental
threat when injected into borehole fluids.

Winter 1996

100
3800 BOPD
0
85

87

89

91

93

Deviation, degree

Two-phase velocity measurements in the


Schlumberger Cambridge Research flow loop. Oil and
water velocity measurements made using the PVL
technique in a laboratory flow loop with two-phase
flow where the water flow rate was maintained constant at 1500 BWPD. The loop was tilted from 85 to
92 degrees and the water and oil velocities measured
for oil flow rates ranging from 750 to 3800 BOPD.
The results show that small deviations from horizontal can cause large changes in the measured fluid
velocities.

51

the casing below X050 m causing the temperature to change faster than the local
geothermal gradient. Above X050 m, the
WFL data revealed flow inside the casing, in
good agreement with the production logging interpretation (right).
The WFL interpretation helped pinpoint
the three-phase production to Zones 2 and
3. Only gas and oil enter the well from
Zone 1. The WFL data show that water, from
below Zone 5, flowed behind the casing.
With a clear understanding of the production problems in the well, the operator
could choose between two remedial treatmentseliminating all water production by
closing Zones 2 and 3, simultaneously cutting potential oil production by a third; or
simply decreasing water cut by repairing the
cement below X050 m.
The next field example shows how a new
PL holdup and velocity imaging tool
helped determine the correct remedial
action for a well on the North Slope,
Alaska, USA operated by ARCO Alaska Inc.
and BP Exploration (next page, left).10
The 49 deviated well, was flowing at
1141 BOPD [181 m3/d] with 82% water cut
at surface and a GOR of 2583 ft3/bbl. Four
zones were originally perforated, and traditional PL interpretation based on density,
velocity and temperature indicated mixed
water and oil production in the lower three
zones, and gas in the top two. For example,
in the lowest perforated zone, the gradiomanometer showed a reduction in fluid
density, usually interpreted as first hydrocarbon entry. Based on traditional PL measurements and interpretation, only this lowest
zone would be produced, and all upper
zones would have been plugged.
A completely different picture emerged
using the recently introduced FloView imaging tool (see, Advantages of Holdup and
Bubble Imaging in Production Logging,
page 54). The FloView water holdup curve
remained at 100% in the lower zone. The
density drop measured by the traditional gradiomanometer probably occurred when the
tool moved from a dense sump fluid lying
below the lowest perforated zone into lighter
water produced from the first set of perforations. Next, the FloView holdup detected a
small hydrocarbon entry in Zone 2, and a
large entry in Zone 3, as seen in the FloView
holdup map.
10. Vittachi A and North RJ: Application of a New
Radial Borehole Fluid Imaging Tool in Production
Logging Highly Deviated Wells, paper SPE 36565,
presented at the SPE Annual Technical Conference
and Exhibition, Denver, Colorado, USA, October
6-9, 1996.

52

Shale
Perforations
Production
Cement
Matrix
Well Sketch
15
in.
-15

-25

WFL GR red
ft/min 100

Theor.Dens
WFL Far blue
1.10
-25 ft/min 100 6.6
gm/cm3
WFL Near green
-25 ft/min 100 Fluid density
0.6
1.10
Fluid Vel
gm/cm3
-25 ft/min 100

Theor. Temp Theor. Pres


235
240 1010 1090
C
psi
Temperature
Pressure
235
240 1010
1090
C
psi

0
0

WFL Water Rates


Gas
Oil
Water
WFL Water Rate
B/D
4000
Downhole Flow Rate
B/D
4000

X025

3
X050

Ou
vel tside
oci
ties

Flo

wo

uts

ide

X075
5

Water flow logs at different depths in a deviated well. Track 1 (left) shows a well sketch
and perforations at each zone. Track 2 shows WFL velocity results. The next three tracks
show PL density, temperature and pressure measurements. Results of flow model analysis
are shown in Track 6 (right). The reconstruction of PL measurements (dashed red) based
on the flow model analysis is shown along with the original (solid black) PL measurements in Track 5. Three detectors were used by the WFL to cover a wide range of flows.
Water velocities inside the casing, derived from the near detector are shown as green circular tadpoles, while the far detector readings are shown in blue and the gamma ray
readings in red. The triangular-shaped tadpoles represent readings for flow outside the
casing. In this display, the 45 angle of the tadpole tails show an upflow in the well.
Downward flow would be indicated by tails pointing 45 downward.

Oilfield Review

FloView
Holdup

Perfs
GR
0

150
API

1:600 ft

0.5

v/v

FloView
Velocity (up)

Oil

1
FloView
Holdup Map

Temperature
218

223

Gradio Density
0.6 gm/cm3 1.1

Conventional PL Results

Gas

FloView
Hydrcarb.
Velocity (down)
0

0.6

ft/min

FloView
Velocity Map

350

1.0 Spinner Velocity 0


v/v

25 ft/min 375

350
ft/min

Gas

Oil

Water
Downhole Flow
Profile
0

Zone

Casing

Water

10,000
B/D

1000

2000

3000

4000

Production, B/D
X800

PL Results with FloView


4

Gas

Zone

Oil
3

Water

X900

1
0

1000

2000

3000

4000

Production, B/D

Comparing production logging techniques. Downhole production from each


zone was measured using conventional
PL techniques and compared with those
from the new FloView imaging technique.
The new technique showed that only
Zone 3 had significant oil production.

3
X1000

Identifying fluid entry. The holdup map in Track 2 and the hydrocarbon velocity map
in Track 4, from an Alaskan well show the first hydrocarbon entry in Zone 3. The center
of each map track represents the high-side of the casing. The difference between the up
(dashed red) and down (solid red) passes of the FloView imaging tool in Track 3 indicates
backflow (shaded grey area where curves cross over) at X900, and hydrocarbon
production (unshaded crossover) in Zones 3 and 4.

In addition, the FloView bubble (or hydrocarbon) velocity map pinpointed the first
significant hydrocarbon entry midway up
Zone 3. The caliper readings, shown as a
casing cross-section profile, supported the
idea that the gradiomanometer interpretation was adversely influenced by changes in
casing diameter between Zones 1 and 3. A
restriction in the casing at X900 ft caused an
increase in both spinner and FloView velocity measurements.
Just above X900 ft, between Zones 3 and
4, there was a reduction in average FloView
bubble velocity. The FloView images
showed a narrow band of hydrocarbon in
this section of the welllow water holdup

Winter 1996

and higher bubble velocity throughout the


top section of the casing. This zone
appeared to have water backflow shown by
comparing an overlay of two passes of the
FloView velocity, one going up the well and
a second traveling downhole. A large separation between the up and down passes was
seen in the region experiencing the water
backflow. The upgoing FloView pass read
higher hydrocarbon velocity than the downgoing pass. This occurred because water
was flowing backwards down the pipe, carrying hydrocarbon bubbles down with it
against the upward motion of the tool. This
abnormal separation in FloView velocities is
an easily recognized flag to spot reverse
flow in the well.

Farther up the well, the opposite occurred.


Starting at Zone 4, the upgoing FloView
pass had a lower hydrocarbon velocity than
the downgoing pass. This occurs because
hydrocarbon bubbles, carried by the
upward flowing water, were moving along
with the upward moving toola sign of significant hydrocarbon entry in Zone 4.
The downhole flow rates and profiles computed from the imaging measurements were
significantly different from those determined
using traditional PL measurements alone.
Flow rates calculated using data from this
new technique were within 8% of actual
production rates (above). Based on these
results, the recommendation to the operator
was to plug off all the zones except Zone 3,
the only significant oil producer.
The overlay techniques shown in this
example can be used as a qualitative
method of identifying zones of hydrocarbon
entry and water backflow.

53

Advantages of Holdup and Bubble Imaging in Production Logging

The 11116-in. FloView production logging tool


makes four independent measurements of bore-

Probe

hole fluids, distributed in different quadrants of the

Ceramic
insulator

pipe cross section (right).

Connector

The self-centralized device uses matchstick-

Casing

sized, electrical probes to measure the resistivity

Conductive
tip

of the wellbore fluidhigh for hydrocarbons and


low for water. The probes are located inside of

Probe holding
bracket
Probe

each of the tools four centralizer blades to protect

Probe

them from damage, and their azimuthal position


within the pipe cross section is measured.
The FloView imager may be run in up to 958-in.
casing. Each probe is sensitive to the local resistivity of the fluid within the pipe and generates a
binary output when their sharp leading edges

FloView images

Water holdup

impinge on droplets of oil or gas in a water-contin-

uous phase, or conversely, water in an oil-continuous phase (next page, left). Assuming the fluids
are distinct and not in an emulsion form, and that
the bubble size is larger than the tip of the probe

0.5

Flow rate
1500 B/D

0.71

0.58

0.48

0.44

(less than 1 mm), both water holdup and bubble


count measurements may be obtained from the
binary output of the

Flow loop photos

probe.1

Water holdup is computed from the fraction of


the time that the probe is conducting, and bubble

80

89

90

91

Deviation
from vertical

count comes from the average frequency of the output. In a water-continuous phase, an increasing
bubble count means an increasing hydrocarbon
velocity, and vice versa in an oil-continuous phase.

Flow-imaging tool and holdup images. The FloView imaging tool has four probes, which map the local water
holdup in the borehole (inset above). FloView images show increasing water holdup as deviation decreases
and correlate well with flow loop photos.

In biphasic fluid flow, the oil or gas holdup may be


obtained from a closure relationship with the water
holdupthe closure relation simply states that the
sum of the holdups of all the phases equals unity.
The probes cannot discriminate oil from gas.
Even in three-phase fluid flow, this device still
yields an accurate water holdup measurement.
Averaged local outputs for holdup and bubble
count are determined for each of the four individual probes. The outputs from each of these probes
are combined to map local stratified holdup.

In a typical two-phase environment, the FloView

cult in highly deviated wells and are impossible in

tool has many advantages over the gradiomano-

horizontal wells because the vertical separation

meter (next page, right). Jetting of producing fluid

between sensor measure points is reduced and the

in front of perforated zones or changes in pipe

measurement loses resolution. Finally, if the flow

diameter because of scale or restrictions have a

velocity is sufficiently high, friction will affect the

venturi pressure effect on gradiomanometer

gradiomanometer response.

response. The gradiomanometer does not measure density directly, but measures the gravitation
pressure gradient with differential sensors over a
known vertical height difference. For this reason,

1. During most field tests, bubble sizes vary between 1


and 5 mm, within the requirements of the probes. Only
at high flow rates (in excess of 2 m/sec [6.5 ft/sec]) are
smaller bubble sizes experienced that might affect the
holdup and bubble-count measurements.

gradiomanometer measurements are more diffi-

54

Oilfield Review

Probe output

Probe

Not
conducting

Conducting

Friction
effects
Third oil
entry

Second
oil
entry

Jetting,
venturi
effects

First oil
entry

Flow

Water
entry

FloView
bubble
count

FloView
holdup

Stagnant
water

Mud

Oil
Time

Gas

Principle of local probe measurement. Oil and gas do not conduct electric
current, but water does. Water holdup is determined by the fraction of time
the probe tip is conducting. Bubble count is determined by counting the
nonconducting cycles.

Winter 1996

Gradio

FloView tool and gradiomanometer comparison in two-phase flow. At the bottom


of the well (middle), there is frequently some mud and dense stagnant water. The
gradiomanometer (right) responds to density change, and will detect the density
decrease above the stagnant fluid, which in many cases might be mistaken for oil
entry. FloView probes do not respond to the water change since both water and
stagnant water are conductive. Therefore, the holdup (left) remains at 100% and
the bubble count stays at zero. The next zone is producing water, typically opposite
perforations. The gradiomanometer detects another density change, and as before,
this change may be misinterpreted as an oil entry, because the produced water is
invariably less dense than the stagnant water. Once again, FloView probes do not
respond to this water change since both waters are conductive. At the first oil entry
in the next zone, the outputs of the FloView probes will indicate less than 100%
water holdup, and the bubble count will start to increase. The gradiomanometer
density will also record the change, if enough oil enters, and the oil density is sufficiently different from the produced water. As the tool passes across additional oil
entries, FloView water holdup will continue to decrease and the bubble count will
increase. The gradiomanometer will also register these oil entries with a decreasing
density, if the oil entries change the mixture density significantly.

55

CPLT Combinable
Production Logging Tool
Pressure and temperature

RST Reservoir Saturation Tool


Oil holdup
Gas indicator

FloView tools
Bubble velocity
Water holdup

Fluid marker
injector

Total flow rate

Gamma ray
detector

CPLT

GR

RST
FloView Plus tool
WFL Water Flow Log
Water velocity
Water holdup
Water flow-rate index

Spinner

PVL Phase Velocity Log


Marker injection for oil
and/or water velocity

The PL Flagship tool string. This composite string consists of the CPLT Combinable Production Logging Tool, an RST module with an
extra gamma ray tool, used for water flow logging and PVL Phase Velocity Logging, a FloView Plus fluid imaging tool, a fluid marker
injector tool used with the PVL, and a total flow rate spinner tool. The two imaging FloView tools are mounted with their probes aligned
for enhanced coverage of the borehole cross section.

Horizontal Wells: The Flagship Project

During 1994, British Petroleum Exploration


Operating Co. Ltd. and Schlumberger Oilfield Services established a joint initiative
The Flagship Projectto develop new
techniques for the diagnosis and treatment
of high-angle and horizontal well production problems.
The diagnosis part of this project involved
development of new PL tools. First, a novel
tool string incorporating sensors targeted at
the stratified flow regimes encountered in
horizontal and near-horizontal wells was
developedcombining the CPLT tool, an
extra gamma ray detector, the RST tool,
FloView Plus tool, fluid marker injector and
a total flow rate spinner tool (above).11 This
equipment is now being used in the North
Sea and the Middle East to make quantitative
flow-rate measurements of oil and water in
cemented and perforated liners, with a longterm goal of being able to measure threephase flow in uncemented liners.
The first application of this tool string was
to resolve flow profiles and monitor movement of OWCs in the Sherwood sandstone
reservoir, in the Wytch Farm field that straddles the coastline of southern England. Using
extended-reach drilling technology, at least
ten onshore wells were drilled with stepouts
of up to 8000 m [26,248 ft] and having
reservoir sections of up to 2700 m [8858 ft].
The wells have electrically submersible
pumps (ESPs) and produce up to 20,000
BOPD [3178 m3/d]. To manage the field, BP
employs production logging on selected
wells to assess flow profiles with respect to

56

Water holdup
Above

0.94

0.88 - 0.93
0.82 - 0.87
0.76 - 0.81
0.71 - 0.75
0.65 - 0.70
0.59 - 0.64
0.53 - 0.58
0.47 - 0.52
0.41 - 0.46
0.35 - 0.40
0.29 - 0.34
0.24 - 0.28
0.18 - 0.23
0.12 - 0.17
0.06 - 0.11
Below

0.5

Average holdup = 0.261

Holdup image from Wytch Farm 1F-18SP well. Multiple positions of the imaging probes
provide a detailed local holdup image. From this image, the local holdup profile is combined with the different phase velocities to determine multiphase fluid-flow rates.

reservoir zones and to monitor the movement of OWCs. This information is used to
determine future well trajectories, optimize
standoff from the OWC and target future
well intervention needs, such as to shut off
water and add secondary perforations.

11. Lenn C, Bamforth S and Jariwala H: Flow Diagnosis


in an Extended Reach Well at the Wytch Farm Oilfield Using a New Tool string Combination Incorporating Novel Production Technology, paper SPE
36580, presented at the SPE Annual Technology Conference and Exhibition, Denver, Colorado, USA,
October 6-9, 1996.
12. Roscoe B: Three-Phase Holdup Determination in
Horizontal Wells Using a Pulsed Neutron Source,
paper SPE 37147, presented at the 1996 SPE International Conference on Horizontal Well Technology,
Calgary, Alberta, Canada, November 18-20, 1996.

Oilfield Review

Winter 1996

Multiphase Holdup Measurements

Gas Holdup Response

Casing
Inelastic Spectrum

YG

Gas

Inelastic N/F ratio

Near and Far count rate


YG = 0.00
YG = 0.33
YG = 0.67
YG = 1.00
Porosity

YG

RST
Tool

Near and Far C/O ratio

oil

oil

ole

tion

YW

eh
or

ma

Water
Energy

C/O Model Response

For

YO

Oxygen

For
wa matio
ter
n

Carbon

Far C/O ratio

Counts

Three Wytch Farm wells were chosen to


evaluate the new Flagship tool stringtwo
with water cut and one a dry-oil producer.
The first water-cut well 1F-18SP was drilled
to a 4450 m [14,600 ft] total depth, with a
horizontal displacement of nearly 3800
m [12,468 ft]. Once the main drainhole was
drilled through the productive section, the
well trajectory was dropped to penetrate the
OWC. This permits future logging of the
OWC as it moves. The reservoir was perforated 33 m [106 ft] above the initial OWC,
giving an initial estimated productivity index
(PI) of 100 B/D/psi.
Production started at 15,000 B/D
[2384 m3/d] dry oil, declining after three
years to a rate of 13,000 B/D [2066 m3/d]
fluid with a 9 to 14% water cut at the time
of logging. This well was selected to test the
new tool string because it had the highest
water cut in the field, penetrated the OWC
and presented the best opportunity for
coiled tubing intervention.
Despite using a revolutionary tool string
for this trial, the logging objectives were typical of any PL job: To determine the source
of water production, identify the oil and
water profile in the well and assess each
zones contribution, and determine any
movement of the OWC in the reservoir.
Analysis of the PL data revealed that the
well was producing fluid along the entire
length of its perforated section. Water production was occurring only in the lowest
perforationsin the toe of the wellpossibly due to coning in a zone of high vertical
permeability, rather than a general movement of the OWC.
The RST-Sigma saturation monitor logs
showed that the OWC had moved up only
10.8 m [35 ft] from its original position. The
independent WFL velocity and the PVL
water-velocity measurements both showed
good agreement with the PL results. In addition, oil-velocity measurements were
obtained from the PVL tool.
Local probes on the FloView tool provided
holdup distribution images of the fluids,
confirming that the flow was stratified (previous page, bottom). In addition, RST-C/O
ratio and borehole salinity from the RSTSigma logs were used for holdup analysis
(see Multiphase Holdup Measurements,
right). 12 All three methodsFloView
Images, C/O ratio, and borehole salinity
provided similar results, confirming trends
or conclusions about holdup analysis. Flow
profiles were computed from the velocity
and holdup measurements for both oil and
water phases.

ate

le w

ho

re
Bo

Near C/O ratio

Multiphase holdup from RST tool. Inelastic spectra (left) lead to carbon-oxygen ratios and near-and-far
detector count rates. The crossplot of near and far C/O ratio responses are determined primarily by oil holdup
YO in the borehole (lower right plot) and oil volume in the formation. The near-to-far inelastic count rate ratio
(upper right plot) primarily depends on the overall borehole density which is related to the borehole gas
holdup, YG.

Multiphase holdup measurements are made with

The RST C/O crossplot response and RST-A

the basic RST C/O measurement, which is usually

inelastic near-to-far count rate ratios are used

used to determine the volume of oil in the forma-

together to determine multiphase fluid holdup.

tion. The carbon and oxygen signals are generated

The inelastic spectra give carbon-oxygen ratios and

by fast neutron inelastic scattering, which leaves

detector count rates. The crossplot near detector

these elements in high-energy excited states that

C/O ratio response is determined primarily by the

decay immediately by gamma ray emission.

oil holdup in the borehole and the far detector C/O

Most carbon-oxygen excitations take place within


15 to 23 cm [6 to 9 in.] of the tool. This means all

response by the oil volume in the formation. The


near-to-far inelastic count rate ratio primarily

C/O measurements are sensitive to the local ele-

depends on the overall borehole density, which is

mental concentrations, and therefore to the relative

related to the borehole gas holdup.2 Two-phase (oil

amount of oil and water holdup in the borehole as

and water) holdup is determined using the crossplot

well as the saturations in the formation (above).

C/O response, while both crossplot and count rate

The RST-A tool has two detectors with one more

ratios are used simultaneously for holdup determi-

and one less sensitive to the borehole environ-

nation in three-phase (oil, water and gas) solutions.

ment by virtue of their spacing from the source.

1. Roscoe B: Three-Phase Holdup Determination in


Horizontal Wells Using a Pulsed Neutron Source,
paper SPE 37147, presented at the 1996 SPE International Conference on Horizontal Well Technology,
Calgary, Alberta, Canada, November 18-20, 1996.
2. An approach to measure borehole gas holdup with a
fullbore backscatter gamma ray density tool can be
found in: Kessler C and Frisch G: New Fullbore Production Logging Sensor Improves the Evaluation of
Production in Deviated and Horizontal Wells, paper
SPE 29815, presented at the 1995 Middle East Oil
Technical Conference, Manama, Bahrain, March
11-14, 1995.

Gamma ray spectra from both detectors lead to


relative elemental carbon and oxygen yields,
which are used to solve simultaneously for the volume of formation oil and the borehole oil holdup.1

57

The data acquisition capability of the tool


string allows most critical parameters to be
determined by alternative independent
methodsfor example, C/O and imaging
holdup data, or WFL and PVL velocity data
supported by spinner measurements
instilling greater confidence in the results.
The new tool string clearly identified all
the water entry points in the well, confirmed
that the downhole flow was stratified, and
proved that water and oil flow rates could
be accurately determined using the new
phase velocity and C/O-based holdup measurements. The upper perforations were producing oil. Oil flow rates derived from the
PVL velocity and C/O holdup, within 500
B/D [80 m 3 /d], were 12,500 B/D
[1986 m3/d]. The water-flow rates derived
from the PVL and WFL measurements,
within 500 B/D, were 3500 B/D [556 m3/d].
In the second water-cut well to be logged
with the PL Flagship tool string, water entry
was found to be not from the toe as before,
but from a nonsealing intersecting fault. The
logs showed that water was being drawn up
through the fault from the OWC.
In the third wella dry-oil producerthe
PVL oil-velocity measurements were tested
against a fullbore spinner flowmeter in the
horizontal drainhole completed with sand
screens. The PVL data matched the spinner
velocity, which functioned effectively in
monophasic production.

Log Inputs

58

Depth Matching

Data Editing
7600

Depth Matching

7700

7800

Fluid Velocity

Calibrations
Spin - rpm

Stacking
Bot. Top

Cable
Speed
Slope Intercept

7750 7700
7800 7750

.21
.22

.02
.03

Calibrations
Initialization

Reconstruction
Flow

Blocking
Flow Model

Velocity Temp

7600

Tool Model

Flow Rate
Solution

7700

Tool
Incoherence

7800

S
o
l
v
e
r

Final Results
Tool
Incoherence
Report and Well Sketch

Tying It All TogetherInterpretation

Traditional PL interpretation for vertical


wells primarily uses density from the gradiomanometer to compute oil and water
holdup, and the averaged measured
flowmeter velocity from the spinner to compute fluid-flow rates using the slip velocity
computed from a fluid model.13 Pressure,
temperature and other data are largely
ignored by conventional PL analysis.
However, such a limited approach is inadequate for most wells. By using all available
production logging data, more complete
answers may be delivered with greater confidence. The BorFlo production logging analyzer is being introduced to do this (above
right). This single interpretation package uses
physical models based on fluid dynamics in
deviated and horizontal boreholes, relating
the physics of fluid flow to the parameters
measured by the PL tools (see Interpreting
Multiphase Flow Measurements in Horizontal Wells, next page). With this interactive
PL interpretation tool, measurements may be
stacked, tool responses calibrated and flowrate solutions determined.

Well and Fluid


Characteristics

Inputs

7600

Gas
Oil
Water

Upper
perfs

7700
7800

Lower
perfs

BorFlo overview. The PL interpretation program allows the engineer to do log stacking,
calibrations and define well and fluid characteristics interactively. The interpretation
matches the PL measurements with those determined by a fluid-flow model based on different flow conditions occurring at each interval.

Multiple measurement of production


parameterssuch as fluid velocities from
spinners, WFL and PVL logging runs, as well
as holdup measurements from imaging tools
and RST logsenable delivery of optimized
solutions to the fluid-flow dynamics. Knowledge of sensor responses allows the optimization to be based on the confidence levels of each logging measurement.
This forward-modeling program tests the
results of different flow conditions, based on
many iterations, to determine the most likely
downhole fluid-flow regime that is consistent with all the borehole geometries, wellbore environment, and observed production
logging and surface measurements.14

13. Slip velocity is the difference between the two-phase


average velocities. For discussion of traditional production log interpretation: Hill AD, reference 2.
14. For example, the Duckler analytical model is used to
determine parameters of the gas/liquid flow regime,
and the volumetric model developed by Choquette
and Piers separates the oil/water regime. For more on
the development and use of the constrained solver
PL interpretation models such as PLGLOB: Torre J,
Roy MM, Suryanarayana G and Crossoaurd P: Go
with the Flow, Middle East Well Evaluation Review
13 (1992): 26-37.

Oilfield Review

Interpreting Multiphase Flow Measurements in Horizontal Wells

A new fluid dynamics-based interpretation model

eter D, deviation angle, , fluid densities o and

called the Stratflo model has been developed to

w, and viscosities, o and w. For example, in

compute oil-water flow rates from logging mea-

terms of the dynamic parameters Vw water veloc-

surements in high-angle and horizontal wells.1 The

ity and Vo oil velocity and Yw water holdup, the

Vo

model depends on basic flow equations, which, in

function can be expressed as

Vw

Tow
Ti

turn, depend on dynamic parameters such as fluid


P in water = P in oil

as well diameter, borehole deviation, and fluid


densities and viscosities. Frictional terms at the

This function is a nonlinear algebraic equation

casing wall are based on monophasic results

and a function of three independent variables.

(right). At the phase interface a simple flat inter-

such as local holdup and velocity measurements

the frictional factor between the two phases has

may be used for two of the necessary input

been developed from flow-loop measurements.

dynamic parameters. With the mass conservation

The model is based on the principle that the

equations, which relate flow rates, velocities and

pressure variation P along the axis of the well in

water holdup, the model can be solved for other

each phase is equal. In steady state, the pressure

combinations of inputs, depending on available

variation in each phase has a hydrostatic compo-

data. Outputs are computed from the flow model

nent, which depends on density and the borehole

and mass-conservation equations using a root-

deviation (the difference in height of the vertical

finding technique.
6000 B/D [953 m3/d] for each phasethe limit

wall for oil Tow and water Tww, and the shear stress

where the simple flat interface starts to degener-

The steady-state model simply sets the pressure in the oil Po equal to the pressure drop in
the water Pw, by defining a function
F = Po Pw = 0.
In this model, the function F depends on dynamic
parameters such as flow rates and holdup, as
well as static parameters, such as flowing diam-

Winter 1996

ate as the mixing layer grows. The model accurately accounts for the variation in holdup at different borehole angles and flow rates (right).
1. Theron BE and Unwin T: Stratified Flow Model and
Interpretation in Horizontal Wells, paper SPE 36560,
presented at the 1996 SPE Annual Technical Conference and Exhibition, Denver, Colorado, USA, October
6-9 1996.

Wall friction (Tw)


Interfacial friction (Ti)
P

Gravity (,devi)

Stratified flow model. The flow model for twophase flow equates the pressure difference due to the
hydrostatic head (which depends on borehole deviation angle ), h, and the wall, Tw, and interfacial, Ti,
friction components for each of the two fluids.

Flow model
Flow model
Flow=800 B/D

1.0

The flow model gives good results up to about

be divided into two parts: the shear stress on the


on the fluid interface Ti.

Pressure Drop

To use the model, readily-measured parameters

face frictional model is assumed. A correlation for

positions), and a frictional component, which can

Tww

F(Vw, Vo, Yw) = 0.

Flow=7000 B/D

0.8

Water holdup

velocities and holdup, and static parameters such

0.6

0.4

0.2

0
87

88

89

90

91

92

Deviation, deg

Measured and predicted holdup variation. Holdup


was measured at different deviations and flow rates
in the Schlumberger Cambridge Research flow loop
and compared with results predicted by the stratified
flow model StratFlo. The results show the rapid variation in holdup with borehole deviation at low flow
rates (red curve), as well as the reduced holdup sensitivity at a high flow rate (yellow curve). The results
are shown for a water cut of 50%.

59

The Outlook

The ongoing development effort in understanding three-phase flow is delivering


resultsincluding detailed gas holdup and
velocity measurementsthat are reshaping
PL services. However, there is still an important flow domain not adequately covered by
todays technologyenvironments where
there is low water holdup and significant
drainhole deviation. Work is under way at
SCR to understand the complex fluid
dynamics, flow instabilities and phase mixing in all regions. This experimentation
together with hydrodynamic modeling will
lead to better future understanding and
management of flow in the borehole (right).
Improved instrumentation and tool technology are also promising faster, more efficient and lower-cost servicessome using
slickline. Other applications will see permanent downhole sensors used for production monitoring. 15 These devices are
rapidly becoming more sophisticated, measuring properties other than temperature
and pressuresuch as hydrocarbons and
phase mixing.
The outlook for production logging is certainly brighter now that it has been at any
time during the last decade. Operators can
look forward not only to a better understanding of their reservoirs, but also to use
of this knowledge for more effectively managing their assets.
RH
15. Baker A, Gaskell J, Jeffery J, Thomas A, Veneruso T,
and Unneland T: Permanent MonitoringLooking
at Lifetime Reservoir Dynamics, Oilfield Review 7,
no. 4 (Winter 1995): 32-46.

Computed 3D droplet-averaged simulations of two-phase flow showing the effects of


shear instabilities. Mapped projections of fluid holdup are shown for horizontal (top) and
vertical (middle) lateral cross section of the borehole and at four positions cutting vertically
across a borehole (bottom). Oil (red) rises due to buoyancy forming an emulsified layer of
oil on the high side of the pipe. The lighter, upper layer flows at a higher velocity than does
the water (blue). This shear flow becomes unstable and an instability occurs that causes
the emulsion of oil to disperse in the water: large eddies mix the two phases up. Then the
process repeats farther up the pipe. Such fluid simulations help scientists test fluid-flow
models under many conditions and design better methods to measure their properties.

Technology Forum
In conjunction with the Schlumberger ClientLink initiative,
Oilfield Review announces its first online technology discussion.
The Production Logging Web-Forum
is an interactive site for comments on this article, inquiries about technology, or open
discussions concerning production logging tools, interpretation and applications.
To access this forum, point your browser to the following URL:
www.connect.slb.com/forums/pl/
If you have problems making the connection, please E-mail:
skipton@slb.com

60

Oilfield Review

Das könnte Ihnen auch gefallen